Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Reinventing Flood Control
Daniel A. Farber1, Robert G. Bea,2 Karlene Roberts,3 Edward Wenk,4 and Kofi Inkabi5
The excuse we have heard from some government officials throughout this
investigation, that Katrina was an unforeseeable ultra-catastrophe, has not only
been demonstrated to have been mistaken, but also misses the point that we
need to be ready for the worst that nature or evil men can throw at us. Powerful
though it was, the most extraordinary thing about Katrina was our lack of
preparedness for a disaster so long predicted. . . . Our purpose and our
obligation now is to move forward to create a structure that brings immediate
improvement and guarantees continual progress.6
I. Introduction
What do the following pairs of accidents have in common?
•
Torrey Canyon tanker (1967) and the Exxon Valdez tanker (1989).
•
The U.S.S. Greeneville (2002) and the U.S.S. San Francisco (2005).
•
The Challenger Space Shuttle (1986) and the Columbia Space Shuttle
(2003).
•
The Piper Alpha Platform (1988) and the Petrobras P36 Platform (2001).
•
Hurricane Betsy (1965) and Hurricane Katrina (2005).
1
Sho Sato Professor and Director, California Center for Environmental Law and Policy, University of California,
Berkeley. This article is an outgrowth of earlier research into the failure of the New Orleans levee project under
NSF Grants No. CMS-0413327 and CMS-0611632, which resulted in the report by R.M. Seed, et al., Investigation
of the Performance of the New Orleans Flood Protection Systems in Hurricane Katrina on August 29, 2005,
available at http://www.ce.berkeley.edu/~new_orleans/.
2
Professor of Civil and Environmental Engineering, University of California, Berkeley.
3
Professor of Organizational Behavior and Industrial Relations, Haas School of Business, University of California,
Berkeley.
4
Professor Emeritus of Engineering and Public Affairs, University of Washington.
5
Ph.D candidate, Civil and Environmental Engineering, University of California, Berkeley.
6
Hurrican Katrina: Recommendations for Reform: Hearing Before the Comm. on Homeland Security and Gov’t
Affairs, 109th Cong. (2006) (statement of Sen. Susan M. Collins, Chairman, Comm. on Homeland Security and
Gov’t Affairs).
1
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
The answer is simple: In each of these cases someone, somewhere, understood that
organizational and system processes were as much the cause of an initial disaster as were
engineering design, construction and maintenance errors.7 In each case, this knowledge failed to
prevent a second disaster from happening in the same industry.
This record suggests that we are doomed to a future in which increasingly complex
organizations fail, causing unnecessary death and injury, large scale economic disruption,
political haggling, and years of rebuilding. Hurricane Katrina exemplified this failure to learn
from the past, and the odds are good that the future will provide yet more tragic examples of this
syndrome. We may be doomed to repeat these failures, even though preventing disasters is
cheaper than recovery. We may be doomed to this future -- despite the fact that we know that
technological failures virtually always occur within the context of management failures, and
despite the growing body of literature that describes management concepts that could reduce
large scale failure.8 But planning for disaster seems to challenge our organizational capacities.9
We are doomed – unless. This article deals with “unless.” The investigation of the New
Orleans levee failure in which all of us participated revealed that the destruction of New Orleans
was a manmade disaster, caused by the organizational structures that designed, built, and
maintained the New Orleans flood control system. Levees failed, not because of the
unavailability of the information needed to build them properly, but because of organizational
inability to act on that information. Our focus here, however, is not on what went wrong: it is on
how to make things go better in the future.
Part II lays the groundwork for understanding the relationship among human error,
7
See, for example, Vaughan, D. (19 ) The Challenger Launch Decision: Risky Technology, Culture, and
Deviance at NASA., Roberts, K.H., and W.H. Moore. (1993) "Bligh Reef Dead Ahead: The Grounding of the
Exxon Valdez." In K. H. Roberts (Ed.), New Challenges to Understanding Organizations. New York: Macmillan,
157-167.
8
E.g., Karlene H. Roberts & Robert G. Bea, Must Accidents Happen: Lessons from High Reliability Organizations,
15 ACAD. MGMT. EXEC. 70 (2001); Karlene H. Roberts & Robert G. Bea, When Systems Fail, 29.3
ORGANIZATIONAL DYNAMICS 179 (2001); SIDNEY DEKKER, THE FIELD GUIDE TO HUMAN ERROR INVESTIGATIONS
(2002); KARL E. WEICK & KATHLEEN M. SUTCLIFFE, MANAGING THE UNEXPECTED: ASSURING HIGH PERFORMANCE
IN AN AGE OF COMPLEXITY (Univ. of Cal., Berkeley 2001).
9
As an example of what doesn’t happen, the National Incident Management System (NIMS) Integration
Center issued this alert:
All federal, state, local, tribal, private sector and non-governmental personnel with a direct role in
emergency management and response must be NIMS and ICS trained. This includes all emergency
services related disciplines such as EMS, hospitals, public health, fire service, law enforcement,
public works/utilities, skilled support personnel, and other emergency management response,
support and volunteer personnel….
Fed. Emergency Mgmt. Agency (FEMA), Our Top Five Most Frequently Asked Questions, NIMS Alert
No. 04-006 (2006), http://www.fema.gov/pdf/emergency/nims/5_common_faqs.pdf.
Another example of the phenomenon discussed in the text: In mid March, 2005, Donald Hiett, Jr, Principal,
Organizational Strategic Solutions Group, was asked by Louisiana State University (LSU) to develop a NIMS
training program directed to the senior executive leadership in New Orleans to take place before June, 2006. On
March 28, 2005 he was informed there was no interest by these officials in taking this training program. Donald
Hiett, Jr. to Karlene H. Roberts, April 13, 2005.
2
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
technology, natural events—and the risk created by the intersection of all three. Part III examines
the failure of political agencies and actors to operate coherently and interdependently, asking the
reader to adopt a new perspective on the U.S. Army Corps of Engineers. Part IV assesses the
prospects for preventing another Katrina. Part V makes a list of specific recommendations.
Where disasters are involved, failing to learn from the past does lead to repeating it. But
we think quite a bit can be learned from the failures that lead to the New Orleans disaster. Some
key lessons concern the Army Corps of Engineers. The Corps needs better funding, an
institutional culture that is open to new data and new engineering practices, peer review of its
flood control projects, and better oversight from Congress and the White House. It also needs to
be networked with state agencies, outside stakeholders such as environmental groups, and other
federal agencies.
II.
Risk, Technology, and Human Errors
The first step in disaster prevention is understanding the role of human and organizational
error in creating and ameliorating risk. People tend to think of risk as a physical phenomenon
stemming from natural events or complex engineering projects such as a nuclear reactor. Such
physical phenomena are a necessary component of risk, but they are only the starting point in
addressing safety concerns. Whether a risk materializes and the extent of the harm it causes are
almost always mediated by human actions. Those actions, in turn, take place inside
organizations with their own histories and cultures. To understand risk, we need to see the
human context as well as the physical events that cause harm. Only then can we begin to
determine the appropriate response to risk. 10 This section lays the groundwork for understanding
the relationship among human errors, technology, natural events—and the risk created by the
intersection of all three.
A. Risks and Technological Delivery Systems
The New Orleans flood offers a good illustration of this relationship. It represented the
technological failure of inadequate flood control measures against a predictable, risky and
potentially lethal event.11 The collapse of the New Orleans Flood Defense System (NOFDS), and
the resulting death and destruction occurred not simply because of the extreme forces of nature,
but because of human and organizational errors (HOE). Indeed, precisely because the
engineering parameters of flood control are so well understood, experts can increasingly focus on
the role of HOE failures in causing disasters. A number of studies of Katrina have concluded that
HOE failures far outnumbered engineering ones as the primary cause of the Katrina disaster.12
Understanding HOE failures involves an assessment of risk. Risk can be defined as a
condition in which either an action or its absence poses threats of socially adverse and sometimes
extreme consequences. Risk derives from acts of nature, from weaknesses of human nature, and
from side effects of technology, all situations that mix complex technical parameters with the
10
For general discussions of risk standards, see Clayton P. Gillette and James E. Krier, Risk, Courts, and
Agencies, 138 U. Pa. L. Rev. 1027 (1980); Cass Sunstein, Which Risks First?, 1997 U. Chi. L. F. 101 (1997).
11
For an overview of the failures in planning the levee system, see JOHN MCQUAID AND MARK
SCHLEIFSTEIN, PATH OF DESTRUCTION: THE DEVASTATION OF NEW ORLEANS AND THE COMING AGE OF
SUPERSTORMS 70-86 (2006). For a summary of the storm’s impact, see DANIEL FARBER AND JIM CHEN, DISASTERS
AND THE LAW: KATRINA AND BEYOND 17 (2006).
12
For the results of our study, see Ray Seed, et al., supra note 1.
3
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
variables of social behavior. Although each risk event is unique, all display commonalities that
permit systemic analysis and management. These recurring properties lead to certain principles.
To begin, the acceptability of a given level of risk cannot be derived from science or
mathematics; it is a social judgment. The spectrum of risk thus includes both the physical world
defined by natural laws, and the human world loaded with beliefs, values, ambiguities and
uncertainties. The physical world follows predictable principles of cause-and-effect. The human
world, by contrast, performs more like an organism, following no set pattern, deeply intertwined
with other actors, and altered by outside events and elements.
Technology mediates the relationship between the social and physical forces that gives
rise to risk. In this context, it is helpful to think of technology as more than hardware; rather, it is
a social system comprising many organizations, synchronized by a web of communications for a
common purpose. It is energized by forces of free market demand, of popular demand for
security and quality of life, and by forces of scientific discovery and innovation. We need to
think in terms of a Technological Delivery System (TDS), a system that applies scientific
knowledge to achieve society’s needs and wants.13 A TDS consists of three primary parts: the
affected public, the governments that represent the public, and the associated industry or
technological enterprise. Inputs to the TDS consist of values, beliefs, technology, and other
resources. Outputs consist of those that are desirable (such as flood protection) and those that
are undesirable (such as floods). Thus, the provision and use of technology mediates risk. In the
case of Katrina, there were critical breakdowns in all parts and levels of the TDS that determined
disaster prepardness.
Put another way, technology acts like an amplifier of human performance. With the water
wheel, the steam engine and the bomb, it amplifies human muscle. With the computer, it
amplifies the human mind and memory. Technology also amplifies social activity, mobility,
quality and length of life. It may also spawn unexpected side-effects which can lead to
complexity, conflict and even chaos. Such effects produce a paradox of risk: technologies are
introduced to defend against the violence of nature, or against human and organizational error,
but often produce unintended and possibly dangerous consequences.
Understanding such risks, and of measures to contain them within safe limits, requires
both hindsight and foresight. The past can illuminate failures and teach lessons for engaging new
issues and threats. The future commands the exercise of foresight, an imaginative preparation of
scenarios stirred by such questions as, “what might happen, if ?” or “what might happen, unless?”
Such inquiries should focus on when disasters might happen, what events will trigger them, who
will suffer, and who is responsible for ameliorating the risk. Modeling becomes essential to
represent all stakeholders and their inter-relationships, including both the private and the public
sectors. The concept of a technology delivery system (TDS) is simply an attempt to model how
the real world works.
B. The Relationship Between Technology Delivery Systems and Political Actors
Federal, state, and local governments are heavily involved in all of the technologies
previously discussed and many more. With waterways, for example, the Army Corps of
Engineers has predominant statutory responsibility. That accords with federal stewardship of
national infrastructure, from roads, shipping channels, harbors, and canals to airplane routes and
13
See Edward Wenk, Making Waves: Engineering, Politics, and the Social Management of Technology
(1995).
4
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
the Internet. The federal budget is constantly challenged to meet a spectrum of different fiscal
demands, the total of which always exceeds Congressional appropriations. Reconciling this
mismatch (often through tradeoffs) becomes the charge of policy-makers at the highest level,
including the President of the United States and the Congress.
Often, focusing on government officials misses a major premise of democratic
governance, the need for transparency and public accountability. This notion is reflected in such
regulatory legislation as the National Environmental Policy Act (NEPA). Section 102(2)(C) of
NEPA14 requires the government to estimate the harm that could result from technological
initiatives, and also to consider the alternatives to accomplish the same goals with less harm.
These environmental impact statements (EIS) are made available for public comment and
possible amendment. This process makes every citizen a part of the governing process, ensuring
they have a say in negotiating the question of “How safe is safe enough?”, and giving them a
stake in the levels of safety and security they desire.
Those who are put in harm’s way must have a voice in what would otherwise be their
involuntary exposure to risk. It is up to the federal agencies to implement this policy, but those
agencies will disregard citizen input unless forced to consider it. Political actors must demand
such accountability. The media can play a role in this principle too. Despite a tendency to
highlight the sensational, the media can enrich popular understandings of risk by proving a
backstory, describing the full cast of stakeholders and explaining the role of each in increasing or
decreasing the risk to life and property of a potential disaster.
Contemporary society demands better protection against threats to life, peace, justice,
health, liberty, lifestyle, private property, and the natural environment. These challenges are not
new, but two things have changed—the increased potency of technology and increased media
coverage. Technology amplifies both human performance and human failures, increasing the
speed in which disasters can happen and their potential harm. At the same time, the media has
sped up too, instantaneously covering worldwide events live, twenty-four hours a day. These two
factors interact after a disaster to pressure political actors for better governance protection and
better governance.
That citizens have higher expectations for disaster preparedness is legitimate, because so
many modern threats are due to human and organizational errors, either in adapting technologies
to meet market demand or in guarding against hazards. If human error is the potential cause of
disasters, then human actors should be able to fix it. The NSF study of New Orleans levee failure
the shows that the Katrina event fits that pattern. Government at all levels failed to provide
security to citizens before and during the catastrophic flooding. Victims are justified in asking
how this pathology of mundane levee failure developed, who is responsible, and how that
knowledge can be applied to prevent a reoccurrence.
C. The Engineering Response to “How Safe is Safe Enough?”
The engineering profession has long understood the need to respond to, and prepare for,
human and organizational errors. Their techniques can illuminate aspects of safety preparedness
and point the way toward designing human organizations capable of minimizing risk.
Engineers have traditionally safeguarded their projects through the technique of
overdesign, a practice of compensating for uncertainties in loading, in materials, in quality of
14
42 U.S.C. 4321 et seq.
5
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
construction and maintenance, etc. This may be accomplished by adopting some multiple of
loading as a margin of safety ranging from 1.4 to 5.0 and even greater. How these margins are
set is of critical importance, especially when tradeoffs with cost, deadlines, or other compelling
factors may compromise the intended reduction of risk.15
This method of safety assurance is more readily applicable to design of mechanisms not
subject to human and organizational errors. The term “errors,” incidentally, is shorthand for a
broad spectrum of individual and societal weaknesses that include ignorance, blunder, folly,
mischief, pride, lack of competence, greed, and hubris. Protecting structures against violence of
nature such as earthquakes, volcanic eruptions, tsunamis, floods, landslides, hurricanes,
pestilence, droughts, and disease may utilize the concept of over-design, based on
meteorological, hydrological, seismic and geophysical data of past extreme events.
Today, two basic aspects of Risk Assessment and Management (RAM) are addressed by
engineers. The first and most prevalent addresses natural (inherent) and technical (modeling,
parametric, state) uncertainties. These factors of safety are derived from two fundamental
approaches: history and explicit evaluations of uncertainties.
The second aspect of risk, not traditionally included explicitly in RAM, is associated with
HOE and frequently termed the “non-technical” aspect. Quality assurance and quality control
are two examples of management processes used to help assure that errors and defects are held to
acceptable levels throughout the life-cycle of the system. But experience developed during the
past two decades clearly shows that many more important approaches and strategies must be
used to successfully address HOE. Reducing HOE is a key goal of our proposals. One
important source of HOE, of course, is failure to properly evaluate the technical risks of a project
– something that was key to the Corps’ failure in New Orleans.
Thus, modern RAM consist of two primary approaches to achieve acceptable safety:
factors of safety, and human and organizational measures. The flooding of New Orleans clearly
shows why both aspects are important. Investigations of the flood control failures show that the
vast majority of water that inundated the city washed through unanticipated breaches in the flood
protection system. These failures, in turn, were firmly rooted in HOE. A “factor of safety” is
generally a design feature that compensates from uncertainties about the potential loading and
also potential errors in construction and maintenance. But the factor of design is meaningless
unless (a) the design takes into account available knowledge, and (b) construction and
maintenance is professionally competent and hence involves a manageable level of risk.
An important example of the irrelevance of numerical factors of safety in the absence of
these conditions is the failure of the 17th Street drainage canal flood wall in New Orleans. Based
on video recordings, reports by residents, and tests of the evolution of the failure, the NSF team
was able to reconstruct what happened early in the morning of August 29th. Winds blew over the
large oak trees growing near the top of the levees that should never have been allowed to grow
there. The trees’ root balls ripped out significant parts of the levee toe connected with
underground seepage, a defect which also should have been corrected earlier. A vertical joint
opened up at the joint between two sections; the joint should not have been weaker than the
segments it connected. At this point, a deep breach opened. As the water continued to rise, the
15
For a good introduction to margins of safety and how they are set, see Henry Petroski, To Engineer is
Human: The Role of Failure in Successful Design 98-106 (1982).
6
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
walls leaned, opening more breaches. In the end, the levee, floodwall and supporting sheet pile
were displaced laterally toward nearby homes by more than forty-five feet.
The fundamental flaw was that the levee was built on unstable soil. Examination of early
maps of the area revealed that bayou tributaries had crossed the canal at the point of failure.
These tributaries were filled with very soft soil that was buried by landfill before the levee was
built. The Corps did not use sufficient boring samples to detect the variations in soil stability.
Further, the Cops built the levee to a very minimum standard of safety (1.3) on experience with
rural rather than urban topography – that is, use of a design strength 1.3 times greater than the
maximum expected loading. This margin of safety was too small given the high human and
economic costs of a levee failure in New Orleans. And there were other flaws. Field tests
performed in 1965 should have warned the Corps that the levee itself was having a dangerous
effect on underlying soil strengths. Even earlier, geologic tests performed by the Corps indicated
the treacherous nature of the soil conditions. And the tops of the floodwalls were almost two feet
lower than projected because of subsidence. Even when surveys in 2004 clearly showed that the
floodwalls were not high enough, no corrective action was taken. Thus, the design was not
based on best available information.
So, the conditions that led to the failure were clearly foreseeable. But they had not been
foreseen. As one engineer in charge of one of the investigations put it, “We did not connect the
dots.” The question, is why not? In our efforts to answer this question, the HOE aspects of the
levee failure became clear to our team. The compounding effects of flawed decisions were
evident. Ignoring early warning signs was pervasive. Breakdowns in communication, rejection
of information, diffusion of responsibility, and in some cases, neglect, were evident. Other
important HOE included failures of foresight, organization, funding, management, and risk
assessment. These were the root causes of the failure of the 17th St. canal floodwall. It might
have helped if the Corps had used a higher numerical margin of safety, but the key problem was
the failure to anticipate and correct human error inside the Corps as well as elsewhere.
Learning from documented failures is a powerful method for reducing risks of repeated
losses. Another method is to learn from close shaves. Many dangerous events fortunately
culminate in only an incident rather than an accident, but the repetition of similar incidents can
serve as early warnings of potential catastrophic failure. Indeed, the practice of systematically
logging and analyzing such “near misses” on the nation’s airways partially accounts for
commercial aviation’s impressive safety record. A system for reporting close encounters of
aircraft was installed decades ago. Anticipating the possibility that airlines might be reluctant to
blow the whistle on themselves, many years ago the Federal Aviation Administration arranged
for NASA to collect incident data and to redact it to protect the privacy of the incident reporter.
NASA also screened reports to identify patterns that could provide early warnings of dangerous
conditions.16 Similar systems are in place for reporting nuclear power plant incidents. Learning
from mistakes allows us to improve designs so as to maintain the desired level of safety.
Giving the growing recognition of human factors in accidents, the NSF team decided to
examine a class of situations where uncommonly high risks were met with startlingly good safety
records. In the Navy, for example, high risks are a part of the daily operations of submarines and
aircraft carriers. Yet accident rates are paradoxically low.17 Careful analysis of these situations
16
[cite]
17
[cite]
7
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
shows that certain qualities of leadership and organizational culture foster the means to prevent
accidents: integrity, a sense of responsibility among all participants, a tolerance by authority
figures for dissent, and consensus on common goals of safe performance. High safety
performance springs from an institutional culture that is bred from the top of the management
pyramid. The most critical element of that culture is mutual trust among all parties.18 Such
institutions are known as High Reliability Organizations (HROs).
HROs are extremely careful in developing and sustaining a culture that rewards all of
those organizational processes that result in extreme reliability and safety. Some of these
processes have to do with training to insure the proper cultural configuration. Some have to do
with communication and decision-making. Decisions often have to be made in a fast-changing
environment. Thus, all contingencies must be considered beforehand to reduce surprises “in the
moment.”
Long experience with military and paramilitary organizations such as first responders
proves the value of rehearsals to reduce risks and control damage. Satisfactory communication is
an essential element of such operation—and rehearsal operations often shown that
communication is the weak link in disaster response. So too do post-accident analyses of real
events, where delays or missteps in the communication of warnings and rescue operations cost
lives.19
Often, catastrophic accidents involve interfaces between groups or organizations. Shift
changes are a notorious example: serious accidents can occur when an outgoing shift fails to
communicate effectively to the next shift about anomalous behavior or emerging problems.
Breakdowns in horizontal communications between organizational units can also serve as
breading grounds for disaster. Some researchers refer to these interstial sources of disaster as the
‘space between’ organizations in which danger lurks.20
As Katrina revealed, the flood control system has been nearly the opposite of a High
Reliability Organization. Instead, it has been plagued by weak leadership, troubled integrity,
group think, inattention to safety, and an inability to learn from experience. Much of the fingerpointing has been at the Corps itself, but as we will see, the organizational flaws are broader.
D. Observations on Risk Management
The most compelling imperative of life is survival. Yet the experience of living teaches
that there is no such thing as zero risk. Some exposures must be tolerated as “normal,” whether
in rush hour traffic or when coping with nature, human fallibility, or the unintended
consequences of technology. To sum up, the context for analyzing the levee failures from
Hurricane Katrina illustrates two simple but important lessons.
First, the design of precautionary measures requires inspired foresight to imagine
alternative futures. All human support systems entail technology, and all technologies project
unintended consequences. Analysis of risk and its control extracts lessons from past failures,
although the most catastrophic events are so rare as to often frustrate projections.
18
19
Karlene H. Roberts, Some Characteristics of One Type of High Reliability Organization, 1 ORG. SCI. 160 (1990).
Id.
20
See, for example,. Cooper, C. and Block, R. (2006) Disaster: Hurricane Katrina and the Failure of Homeland
Security, New York: Times Books Henry Holt.
8
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Organizational learning is a critical component of disaster prevention. To prevent
devastating accidents, organizations need to make learning a high priority; they must be alert for
minor unexpected events that may be harbingers of larger risks; and they must react decisively
when mistakes occur to reform the systems which made the mistakes possible. None of this
comes very naturally to organizations, which often find it easier to perfect routines, optimizing
their behavior under “normal” conditions but creating little capacity to detect or respond to the
abnormal. Research on organizational learning finds that practices and routines in organizations
develop incrementally through feedback from the organization’s environment. Organizations
generally tend to be inert, adapting less than perfectly to, and falling in and out of alignment
with, their environments.21
This stagnation is especially dangerous for organizations that deal with major
emergencies such as floods, fires, and other natural and manmade disasters. Organizations that
await major failures before adapting tend to enter crisis mode and find learning and response
even more difficult.22 For example, following the demise of the space shuttle Challenger, NASA
faced political pressures, inertia, and resource constraints that expedited some organizational
changes but made other structural and cultural adjustments more difficult.23 Furthermore, in the
absence of a significant environmental change or destabilizing event, lessons learned in
organizations often tend to be forgotten or misapplied.24
Even worse, because of the infrequency with which major disasters occur, trial and error
organizational learning processes may lead organizational members to forget lessons from past
disasters. Levitt and March argue that in the case of disaster preparedness, trial and error
processes lead to “pernicious learning” – organizational leaders conclude that resources
designated for disaster preparedness are idle and should be applied elsewhere.25 Disaster
preparation calls for a different form of learning in which organizations draw on not only their
own experiences, but also those of other organizations. Such network effects exist for a variety
of learning processes.26
21
RICHARD R. NELSON & SIDNEY G. WINTER, AN EVOLUTIONARY THEORY OF ECONOMIC CHANGE (1982). See also
Bradbury, H., and B.M.B. Lichtenstein (2000) Relationality in organizational research: Exploring the space
between. Organization Science, 11, 551-564. Heath C. and N. Staudenmayer (2000) Coordination neglect: How lay
theories of organizing complicate coordination in organizations. In B.M. Staw and R.I Sutton (Eds), Research in
Organizational Behavior, 22, 153-191. Elsevier Science.
.
22
Roberts, K.H., Madsen, P., and Desai, V. (2005) “The Space Between in Space Transportation: A Relational
Analysis of the Failure of STS 107,” In M. Farjoun and W. Starbuck (Eds). Organization at the Limit: Lessons from
the Columbia Disaster, Malden, MA: Blackwell, 81- 98
23
HOWARD E. MCCURDY, INSIDE NASA: HIGH TECHNOLOGY AND ORGANIZATIONAL CHANGE IN THE U.S. SPACE
PROGRAM (1993).
24
Pabli Martin De Holan & Nelson Phillips, Remembrance of Things Past? The Dynamics of Organizational
Forgetting, 50 MGMT SCI. 1603 (2004); James G. March et al., Learning from Samples of One or Fewer, 2.1 ORG.
SCI. 1 (1991).
25
Barbara Levitt & James G. March, Organizational Learning, 14 ANN. REV. SOCIOL. 319 (1988).
26
See, e.g., Linda Argote et al., The Persistence and Transfer of Learning in Industrial Settings, 36.2 MGMT SCI.
140 (1990); Joel A.C. Baum & Paul Ingram, Survival-Enhancing Learning in the Manhattan Hotel Industry, 18981980. 44.7 MGMT. SCI., 996 (1998); Christine M. Beckman & Pamela R. Haunschild, Network Learning: The
Effects of Partners’ Heterogenity of Experience on Corporate Acquisitions, 47.1 ADMIN. SCI. Q. 92 (2002).
9
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Over the past few decades, scholars from many disciplines have advocated relational or
systems approaches, as opposed to reductionist approaches that study particular events and
entities in isolation.27 Taking a relational approach will help us identify and examine learning
processes as they affect and are influenced by organizations responding to major catastrophes.
The issues we discuss may occur at several different levels in organizations – the interpersonal
level, the sub-unit level, or the inter-organizational level. Fortunately, we have learned a great
deal about how to overcome these organizational barriers. What is needed is to instill
“mindfulness” toward risks.
Second, risk management is inherently political, in the sense that it involves society’s
judgments about the acceptability of risk. Different cultures have different risk tolerances,
including attitudes distinguishing voluntary from involuntary risk. In a democratic society, the
decision of how much risk is enough must ultimately be made through the political process.
Tradeoffs are inevitable between short- and long-range events and consequences,
between safety and cost, between special interests and social interests, between who wins and
who loses, and who decides. Society embraces a spectrum of values that often conflict, as with
the goals of efficiency in the private sector and of sustainability and social justice in the public.
Key decisions regarding citizen safety and security are made by government through public
policies to manage risk. These policies dominate the legislative agenda. This mandate imposes a
heavy burden on the President and on Congress, both bodies requiring access to authentic and
immediate information. In our democracy, this authority should flow from citizens following the
principle that those who govern do so at the informed consent of the governed. Making decisions
and implementing them necessarily draws political capital, meaning the decisions ultimately
come down to political power and political will.
These issues illustrate the anatomy of risk and the complexity of its management. They
sound a wake-up call for those responsible for risk management to pursue a deeper
understanding of risk—and for those attentive but exposed citizens to seek a voice in the
decision-making process.
III. Organizational Failures and the Katrina Disaster
The risk awareness and management strategies described in Part II must ultimately be
enacted by political agencies and actors who are, after all, charged with protecting the country.
Yet the magnitude of organizational failures during Katrina suggests the state and federal
governments are unprepared for this task. A major part of the problem is that first responder
organizations, do not operate in a vacuum, but are nestled within a large number of organizations
that should be interdependent with one another. This friction leads to failures. Indeed, the
relationships among political actors and the Corps, among the Corps and other federal agencies,
and within the Corps itself can all serve to preclude or frustrate the types of risk management
strategies described in Part II.
27
JAMES GRIER MILLER, LIVING SYSTEMS (McGraw Hill 1972); FRED A. WOLF, TAKING THE QUANTUM LEAP: THE
NEW PHYSICS FOR NONSCIENTISTS (1980); William E. Kastenberg et al.,Shifting the Paradigm of Risk Analysis,
Proceedings of the International Seminar on Nuclear War and Planetary Emergencies, 29th Session, Erice, Italy,
May 10-15.
10
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
The social science literature addresses this problem by using such concepts as
interstices,28 “interdependencies”29 or the “space between.”30 Failure to consider the processes
that operate both within any one unit and across multiple units is failure to be ready for the next
large scale catastrophe. This discussion focuses on that context, and asks the reader to take a new
perspective on the Corps of Engineers.
Failure of the New Orleans Flood Defense System (NOFDS) was not caused by an
overwhelming extreme natural event (hurricane wind, waves, currents, surge, etc.). While
portions of the NOFDS were overtopped by Hurricane Katrina's surge and waves, our studies
indicate that the vast majority of the flooding came from unanticipated and unintended breaches
in the levees (many adjacent to other structures), failures in the floodwalls, and water entering
through gaps (floodgates not in place) or low spots (due to subsidence and incomplete sections)
in the NOFDS. The roots of these unanticipated and unintended developments were firmly
embedded in Technology Delivery System flaws and malfunctions; failures of organizations and
institutions; and their resource allocation processes.
We identified eight categories of TDS malfunctions that played primary roles in the
failure of the NOFDS:
1) Failures of foresight: Catastrophic flooding of the greater New Orleans area due to
surge from an intense hurricane was predicted for several decades. The consequences observed
in the wake of hurricane Katrina were also predicted. The hazards associated with the NOFDS
were not recognized, defensive measures identified and prioritized, and effective action was not
mobilized to effectively deal with the hazards.
2) Failures of organization: The roots of the failure of the NOFDS are firmly embedded
in flawed organizational systems. The organizations lacked centralized and focused
responsibility and authority for providing adequate flood protection. There were dramatic and
pervasive failures in management represented in ineffective and inefficient planning, organizing,
leading, and controlling to achieve desirable quality and reliability in the NOFDS. There were
extensive and persistent failures to demonstrate initiative, imagination, leadership, cooperation,
and management.
3) Failures of funding: The failure of the NOFDS resulted from inadequate provision of
resources based primarily on recommendations provided by the Corps followed by failure of the
federal and state governments to fund badly needed improvements once limitations were
recognized. In several instances, State agencies pressured for 'lower cost' solutions not realizing
that these solutions would result in lowering the quality and reliability of the NOFDS. There
were important deficiencies in the cost-benefit analyses used to justify the levels of protection
and their continued improvement as knowledge and technology advanced.
28
Martha Grabowski & Karlene H. Roberts, Risk Mitigation in Virtual Organizations, 10 ORG. SCI. 704 (1999).
29
Chip Heath & Nancy Staudenmayer, Coordination Neglect: How Lay Theories of Organizing Complicate
Coordination in Organizations. in 22 RESEARCH IN ORGANIZATIONAL BEHAVIOR 153 (Barry M. Staw & Robert I.
Sutton eds., Elsevier 2000).
30
Hilary Bradbury & Benyamin M. Bergman Lichtenstein, Relationality in Organizational Research: Exploring the
Space Between, 11 ORG. SCI. 551 (2000); MARTIN BUBER, I AND THOU (Walter Kaufman trans., Scribner’s Sons
1970).
11
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
4) Failures of diligence: Forty years after the devastating flooding caused by Hurricane
Betsy, the flood protection system authorized in 1965 and founded on the Standard Project
Hurricane (SPH) was not completed. The concept and application of the SPH was recognized to
be seriously flawed, yet there were no adjustments made to the system before Katrina struck.
Early warning signs of deficiencies and flaws persisted throughout development of the different
components that comprised the NOFDS and these signs were not adequately evaluated and acted
upon.
5) Failures of trade-offs: The history of this system was marked by a series of flawed
decisions and trade-offs that proved to be fatal to the ability of the system to perform adequately.
Compromises in the ability of this system to perform adequately started with the decisions
regarding the fundamental design criteria for the development of the system, then were
propagated through time as alternatives for the system were evaluated and engineered. Design,
construction, operation, and maintenance of the system in a piecemeal fashion allowed the
introduction of additional flaws and defects. Efficiency was traded for quality, reliability, and
effectiveness. Superiority in provision of an adequate NOFDS was traded for mediocrity and
getting along.
6) Failures of management: Requirements imposed on the Corps of Engineers by
Congress, the White House, State and local agencies, and the general public have changed
dramatically during the past three decades. Defense, re-construction, maintenance, waste
disposal, recreational, emergency response, ecological restoration have served to divert attention
from flood control. Public and Congressional pressures to reduce backlogs of approved projects,
improve project and organizational efficiency (down-sizing, out-sourcing), address
environmental impacts and develop appropriations for projects have served to divert attention
from engineering quality and reliability of flood control. Engineering technology leadership,
competency, expertise, research, and development capabilities appear to have been sacrificed for
improvements in project planning and controlling.
7) Failures of synthesis: While individual parts of a complex system can be adequate,
when these parts are joined together to form an interactive-interdependent- adaptive system,
unforeseen failure modes can be expected to develop. These unforeseen, but foreseeable, failure
modes did develop in the NOFDS during hurricane Katrina. It is evident that insufficient
attention was given to creation of an integrated series of components to provide a reliable
NOFDS. Synthesis was subverted to decomposition. As a result, many failures developed at
interfaces in the NOFDS.
8) Failures of risk assessment and management: The risks (likelihoods and
consequences) associated with hurricane surge and wave induced flooding were seriously
underestimated. There was inadequate recognition of the primary contributors to the likelihoods
and consequences of catastrophic flooding. Sufficient defensive measures to counteract and
mitigate these uncertainties were not deployed. Factors of safety used in design of the primary
elements in the NOFDS were insufficient. Quality assurance and control measures invoked
during the life of the system failed to disclose critical flaws in the system. Inappropriate use was
made of existing engineering technology available to design, construct, operate, and maintain a
NOFDS that would have acceptable quality and reliability. Deficient risk management methods
were used to allocate resources and impel action to properly manage risks. Risk management
12
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
failed to employ continuing improvement, monitoring, assessment, and modifications in means
and methods which were discovered to be ineffective.
As Houck observes:
So What Do We Do? Here is what we know. It is not just the tire, it's the car. And
it's not just the car, it's the driver. Nothing in the system has made a numero uno
priority either of protecting New Orleans from hurricanes or to restoring or even
hanging onto -- the Louisiana coast. We have a flood control program, a
navigation program, a permitting program, a coastal management program, a
flood insurance program, a coastal restoration program - just for openers - and
they do not talk to each other. They are riddled with conflicts, basically headless,
basically goal-less, weakened by compromises and refuse outright to deal with
first causes and first needs.31
The key phrases here are “and they do not talk to each other” and “[t]hey are riddled with
conflicts, headless, basically goal-less…and refuse outright to deal with first causes….”
In reaction to the organizational liquefaction that developed during Hurricane Katrina, the
Senate Committee on Homeland Security and Governmental Affairs recommended:
The Corps and local levee sponsors should immediately clarify and memorialize
responsibilities and procedures for the turn-over of projects to local sponsors, and
for operations and maintenance, including, but not limited to procedures for the
repair or correction of levee conditions that reduce the level of protection below
the original design level (due to subsidence or other factors) and also emergency
response. It must always be clear - to all parties involved - which entity is
ultimately in charge of each state of each project. The Corps should also provide
real-time information to the public on the level of protection afforded by the levee
system. A mechanism should be included for the public to report potential
problems and provide general feedback to the Corps.32
A. The Office of the President, the Congress, and the Corps
Other things happen at interstices. The interfaces between the Corps and its political
masters, Congress and the President, are particularly troublesome. The primary interactions
between these entities involve funding. During the preparation of the NSF Report, we gathered
information about funding requests from the President and ultimate Congressional budgeting.
Figure 1 shows the Presidential and Congressional budget requests and Congressional
recommendations for Corps of Engineer funding for 1975 through 2005 for the hurricane flood
defense projects of Lake Pontchartrain and the vicinity.
31
Oliver A. Houck, Can We Save New Orleans?, 19 TUL. ENVTL L.J. 1 (2006).
32
Senate Comm. on Homeland Security and Governmental Affairs, U.S. Senate, Hurricane Katrina: A
Nation Still Unprepared (2006).
13
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Dollars in Thousands
100,000
80,000
Congressional
Budget Request
60,000
40,000
Presidential
Budget Request
20,000
1975
1980
1985
1990
1995
2000
2005
Fiscal Year
Figure 13.1: Lake Pontchartrain and Vicinity Project Construction Appropriations Over the
Past 30 years [2005 dollars]; President’s Budget Request (grey) and the Amount
Recommended by Congress (black).
Several inferences can be gleaned from this information. First, it appears that while the
president was trying to reduce Corps funding, Congress was trying to protect it. With only 60%
of the Lake Pontchartrain projects complete as of 2005 (40 years after authorization), it may be
that Congress, in its wisdom, decided to fund only what it thought needed to be completed. The
graph shows other interesting issues about interdependencies. The Corps of Engineers is
interdependent with both the Office of the President of the United States and Congress. But
because Congress has funded the Corps at higher levels than those recommended by the
President for the last decade, Congress wields enormous power over the Corps, often pressuring
it to fund new large projects in legislators’ districts. As a result, two things happen: First, the
Corps defers needed maintenance projects and instead focuses on the new projects demanded by
Congress. Second, to preserve its budget, the Corps has to devote time to currying favor with
Congress. Currying favor with Congress is not supposed to be a main task of the Corps.
Yet another interesting hypothesis can be derived from these data. When multi-year
projects are funded annually, an interesting dilemma is created for the funded organizations. The
funding oscillation level is at one level, but organizations struggling under that oscillation
oscillate at a higher frequency. It is hypothesized that this is because the funded organization
operates under a considerable amount of ambiguity and uncertainty. This suggests that the
unpredictability of the Congressional process creates unintended and negative consequences for
its funded agencies. The processes and responses to them are both schizophrenic.
This is almost surely the same as the case for NASA. The Columbia Accident
Investigation Board (CAIB) report said:
The White House and Congress must recognize the role of their decisions in this
accident and take responsibility for safety in the future.… Leaders create culture.
It is their responsibility to change it.… The past decisions of national leaders – the
White House, Congress, and NASA Headquarters – set the Columbia accident in
14
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
motion by creating resource and schedule strains that compromised the principles
of a high risk technology organization.33
Diane Vaughan reports that both economic strain and schedule pressure still exist at
NASA. She notes that it is unclear how the conflict between NASA’s goals and the constraints
upon achieving them will be resolved, but that one lesson from Challenger and Columbia is that
system effects tend to reproduce.34 This also happens to military installations every time a Base
Reallocation and Closing (BRAC) list is formed. From the day of its publication until the day of
decisions, the installations on this list spend considerable time trying to get off the list,
distracting them from their principle tasks.
In the Katrina case, will Congress and the Office of the President take a sweeping look at
their own behaviors in concert with those of the Corps of Engineers? They probably will not
because there is not yet a strong incentive for them to do so. One incentive might be that the cost
of cleanup is always more than the cost of prevention. Money is not limitless. But since we’ve
observed many costly past disasters that were not prevented, and many instances in which they
could have been mitigated or prevented, the reality is they probably will do nothing. Thus, the
challenge is to find incentives that will encourage disaster-prevention and emergency-response
organizations and actors, from the President on down, to examine their own organizational
skeletons, muscle, and flesh, as well as to look at the “spaces between.”
B. Other External Interstices for the Corps
Three additional sorts of interfaces between the Corps and its constituents are important.
The first are the interfaces between the Corps and other federal agencies that occur during and
after an emergency, and that are mandated by Emergency Support Function # 3 of the National
Response Plan, a federal-government-wide master plan for emergencies designed and
administered by the Department of Homeland Security”.35
ESF #3 is structured to provide public works and engineering-related support for
the changing requirement of domestic incident management to include
preparedness, prevention, response, recovery and mitigation actions. Activities
within the scope of this function include conducting pre- and post-incident
assessments of the public works and infrastructure; executing emergency contract
support for life-saving and life-sustaining services; providing technical assistance
to include engineering expertise, construction management, and contracting and
real estate services; providing emergency repair of damaged infrastructure and
critical facilities; and implementing and managing the DHS/Emergency
Preparedness and Response/Federal Emergency Management Agency
(DHS/EPR/FEMA) Public Assistance Program and other recovery programs.36
33
COLUMBIA ACCIDENT INVESTIGATION BD., NAT’L. AERONAUTICS
ACCIDENT INVESTIGATION REPORT (2003).
AND
SPACE ADMIN. (NASA), COLUMBIA
34
Diane Vaughan, Systems Effects: On Slippery Slopes, Repeating Negative Patterns and Learning from Mistakes,
in ORGANIZATION AT THE LIMIT: LESSONS FROM THE COLUMBIA DISASTER 41 (William H. Starbuck & Moshe
Farjoun eds., Blackwell 2005).
35
For further information about the NRP, see Farber & Chen, supra note x, at 53-108.
36
DEP’T OF HOMELAND SEC.,NATIONAL RESPONSE PLAN (2004).
15
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
In responding to an emergency, the Corps can draw on the resources of 15 federal
government agencies to accomplish these goals. In addition, state, local and tribal governments
are “fully and consistently integrated into EFS #3 activities.”37 All of this occurs, of course,
when an incident or potential incident overwhelms state, local, and tribal capabilities.
The NRP concept of operations states that the DOD/USACE is the primary agency for
providing ESF #3 technical assistance. It further states that close coordination is to be maintained
with federal, state, local, and tribal officials to determine the potential support needs. In addition,
it spells out the organizational structures for providing support, naming the Interagency Incident
Management Group (IIMG) as the resource for providing on-call subject-matter experts to
support IIMG activities.
Regional and field level mechanisms of support are clearly defined. ESF #3 activities are
also spelled out and include such processes as:
[C]oordination and support of infrastructure risk and vulnerability assessments,
participation in pre-incident activities, such as pre-positioning assessment
teams,… participation in post-incident assessments of public works and
infrastructures to help determine critical needs and potential workloads,
implementation of structural; and non structural mitigation measures, including
deploying protective measures to minimize adverse effects or fully protect
resources, prior to an incident.38
In the wake of Hurricane Katrina, neither the Corps nor any other agency was fully
successful in rolling out the NRP. If the integration required by this plan is too difficult for
agencies to implement, then it is the duty of the agencies and their oversight agencies (e.g.,
DOD, DHS, HHS, etc.) to indicate this and to develop strategies to revise the NRP to create a
workable plan and document. Lee Clarke discusses at length “fantasy plans” that look a lot like
the NRP.39 Thus, a last word on integration across agencies:
From the vantage of preparedness, the failed response to Hurricane Katrina did
not undercut the utility of “all-hazards” planning. Rather, it pointed to problems
of implementation and coordination. This suggests that in the aftermath of the
event, we are likely see the redirection and intensification of already-developed
preparedness techniques rather than a broad rethinking of the security question.40
Given our experiences with accident response, it is this team’s conclusion that, without
substantial leadership and reorganization, neither comprehensive technical nor social reforms
will likely soon be developed to address future natural or man-made catastrophes.
The second set of interfaces that need to be considered are those created by the Corps’
need to “outsource” many functions (e.g., the hiring of outside, private firms and/or individuals
to perform work, including engineering design and construction.) The federal government has
imposed this requirement on the Corps through the White House Office of Management and
Budget and through Congressional actions. In reviewing this mandate with current Corps of
37
Id.
38
Id.
39
LEE CLARKE, MISSION IMPROBABLE: USING FANTASY DOCUMENTS TO TAME DISASTER (Univ. of Chicago 1999).
40
Andrew Lakoff, From Disaster to Catastrophe: The Limits of Preparedness, Understanding Katrina: Perspectives
from the Social Sciences, http://understandingkatrina.ssrc.org/Lakoff (last visited May 15, 2006).
16
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Engineers personnel in multiple settings and in private briefings, our team has concluded that
outsourcing and partnership efforts have threatened the Corps’ key engineering competencies
(practicing, research, development). These competencies were sacrificed under pressures to
outsource, to improve project management, and to develop environmental restoration and
mitigation capabilities, all within a finite overall budget and resources.
This is not to say that such partnering holds only disadvantages. On the contrary, crucial
operational benefits can accrue from partnering. An agency can learn new concepts from
partners, perhaps through access to best-of-class processes. Partnered competitors perhaps can
learn technology secrets from one another. When industry benchmarks aren’t well known,
partnering with a competitor can offer insights on an organization’s productivity, quality, and
efficiency.
But there are also obvious disadvantages to partnering. Lack of control is a critical
disadvantage. The demise of ValuJet, for example, happened because the company outsourced
cargo handling to a company it had no control over in terms of quality standards. In another form
of outsourcing, competitors learn from each others’ operations, which may be detrimental to one
or more partners. Mutual trust is a chronic problem. A new company board for one of the
partners may not approve of the other partner. The strategic aims of partners may change midstream, causing failure. These are just some of the reasons for outsourcing failures.41 The Corps
needs to examine its partner relationships, asking if it has lost too much.
One of the Corps’ sister agencies in times of chaos, FEMA, has also created problems by
outsourcing its disaster response efforts.
For example, when the Nisqually earthquake struck the Puget Sound area in 2001,
homes that had been retrofitted for earthquakes and schools with FEMA funds
were protected from high-impact structural hazards. The day of that quake was
also the day that the new president, G. W. Bush, chose to announce that Project
Impact would be discontinued. Funds for mitigation were cut in half, and those
for Louisiana were rejected. Disaster management was being privatized, with the
person who was to be promoted to head the agency, Michael Brown, saying at a
conference in 2001, “The general idea—that the business of government is not to
provide services, but to make sure that they are provided—seems self-evident to
me.” The administration tried to cut federal contribution for large-scale natural
disaster expenditures from 75 percent to 50 percent.42
D. The Corps’ Internal Interstices
Two other organizational processes also result in lost institutional memory and loss of
control. They are downsizing and retirements. Figure 2 shows that the Corps is also losing
employees through retirements.
41
Karlene H. Roberts & Daniel Wong, Outsourcing and Temporary Workers, in ENCYCLOPEDIA OF
INDUSTRIAL/ORGANIZATIONAL PSYCHOLOGY (Steven Rogelberg ed., Sage 2006).
42
Charles Perrow, Using Organizations: The Case of FEMA, Social Science Research Council Forum,
http://understandingkartina.ssrc.org (last visited Apr. 21, 2006)(internal references omitted).
17
Farber et al., Reinventing Flood Control
Tulane Law Review
Figure 2:
12/10/2006
Human Capital Planning Projected Retirement (the Corps 2002).
In 2002, between 35 and 40 percent of architecture and engineering work was outsourced
to private firms, while all construction projects were outsourced.43 The simultaneous operation
of the three processes (outsourcing, downsizing and retirement) has been and will be disasters for
the Corps. Retirements, downsizing, and outsourcing are interdependent in terms of the problems
they cause for organizations. Again, the causes are probably buried in not only the Corps
activities, but in the Corps' relationships with its external constituencies.
New approaches to organizational failure examine the degree to which organizations are
internally stove-piped. Figure 3 shows that the Corps’ organizational structure might lend itself
to this. It appears that regions and districts act autonomously to a large extent.
43
United States Army Corps of Engineers, USACE Strategic Workforce Planning, USACE Command Brief,
<http://www7.nationalacademies.org/ffc/usace.pdf> (Mar. 30, 2006).
18
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Washington,
Headquarters
Division A
District A1
District AX
Division X
District X1
District XX
Figure 3: Conceptual Organizational Chart of the U.S. Army Corps of Engineers Civil
Works Program.44
In addition Houck observes:
…restoring coastal Louisiana is a national issue and will require remedies beyond
this state. We lie at the receiving end of a large watershed, and some of what we
need has been turned off and other stuff that is hurting us has been turned on. The
Corps districts need to talk to each other. The EPA has to step up to the plate,
upstream states have to change some habits too. If the nation’s taxpayers are
going to be asked to spend more money than America spent on the Marshall Plan
to fix all of post-war Europe, then they have a right to expect a national effort.45
McCurdy discusses how stove-piping existed when NASA was created. Today, the
adverse results of NASA’s stove-piping are excessive unit independence, specialization, and
neglect of mutual coordination in a situation that should be characterized by just the opposite.46
All in all, the Corps’ ability to do its job has been organizationally handicapped. It has
lost engineering and research and development capabilities; it has lost its ability to maintain old
projects; it fails to be appropriately interdependent with various constituencies; and it fails to act
effectively on issues of internal interdependence. And, it cannot get well on its own.
44
The Corps Civil Works Program is composed of 8 Divisions and 38 subordinate districts. Prime Power, ERDC,
Centers, and FOAs are not shown for clarity. In addition, a 9th provisional division with four districts was activated
January 25, 2004, to oversee operations in Iraq and Afghanistan. A more complete organizational chart is available
in USACE 2012 – Appendix G, Resource Analysis, page 1.
45
Houck, supra note 32 .
46
Karlene H. Roberts, et al., The Space Between in Space Transportation: A Relational Analysis of the Failure of
STS 107, in ORGANIZATION AT THE LIMIT: LESSONS FROM THE COLUMBIA DISASTER 81 (William H. Starbuck and
Moshe Farjoun eds., Blackwell 2005).
19
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
IV. Preventing the Next Katrina
In virtually all human affairs, risk is normal. But the consequences of neglect may be
grave. As we indicated in the beginning of this article, we are skeptical that those with the power
and resources to prevent the next Katrina will take the steps necessary to do so.
Nonetheless, from our larger discussion about defining risk and safety in the context of
human actors and organizations, three recommendations emerged. First, decision-making
responsibility at the top of the authority structure should be clarified and strengthened to enhance
vigilance and the management of all modes of risk. Second, Congress should hire additional
technical staff to ensure federal agencies have enough resources, to manage risk, and to monitor
the performance of the Executive Branch in its duties of care. Third, state and local governments
should authorize new processes that would foster informed consent and dissent, provide early
warnings in disaster-prone areas, and ensure that citizens at risk have access to information
regarding their exposure and opportunities to participate in governance.
One central purpose should animate all the entities involved, separately and in tandem.
They should address the question, “How Safe is Safe Enough?” That investigation demands
foresight in the spirit of the injunction, “Without vision, the people perish.”
In addition to this larger purview, specific attention needs to be given to the Corps and
the organizations with which it is interdependent. We know a great deal about how to fix
problems of this nature, and there are growing bodies of engineering, legal, public policy,
organizational, and other literatures that address such issues. There is also a growing pool of
experts from different areas who know how to talk about such issues. The problem is that
stakeholders have huge incentives not to pay any attention to such expertise. They are no more
likely to address these issue than they were likely to prevent the Challenger problem from
becoming the Columbia problem, or the Betsy problem from becoming the Katrina problem.
Fixing these problems – preventing HOE by bridging interstices -- will require a set of
processes that affected stakeholders do not want to engage in, despite their interdependencies:
•
They must come together to fashion clear and consistent goals in a politically complex and
charged world.
•
They must be willing to spend many years addressing such problems even as shifting
political incentives lead to policy attention spans best measured in weeks or hours.
•
Agencies must work together and trust one another.
•
They must recognize their interdependencies as well as the interdisciplinary nature of their
problems.
•
They must be willing to spend money and make recipients of that money accountable for
their spending.
•
They must develop oversight programs and agencies with real teeth.
All of this is easier said than done. We offer some suggestions about initial steps to revitalizing
the federal flood control program.
A. Re-Engineering the Corps
20
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Fixing the Corps’s technical problems will have only limited impact unless we also fix
the organizational problems. The Corps must strive to become a High Reliability Organization.
We have four recommendations that would go a long way toward repairing the Corp’s ability to
design and build effective flood control projects: (1) rebuild the Corps’s engineering and R&D
capability, (2) restructure the federal/state relationship in flood control, (3) develop a National
Flood Defense Authority, and (4) create effective disaster planning. These recommendations are
designed to break down the gaps to the flow of expertise and information within the Corps itself
and between the Corps and outside experts and constituencies.
Three years before Katrina, the National Research Council concluded that the “Corps’
more complex planning studies should be subjected to independent review by objective, expert
panels.”47 This is an obvious point – which makes it all the more urgent to implement. In theory,
independent viewpoints and expertise could be brought to bear within an internal Corps review
process, and that would probably happen in an HRO. But the Corps lacks sufficient technical
expertise as well as the institutional culture of constant internal criticism and learning needed for
such an internal review process to be effective. The prospect of facing independent outside
review can actually help an organization improve its own internal culture in this regard.
Although the need for independent project review has been apparent for years, none of the past
proposals have yet been implemented.48 There is, however, some hope that congressional action
on this issue will be forthcoming.
The major effort to reform the Army Corps of Engineers currently resides in the Senate
version of the Water Resources Development Act (S. 728).49 Congress normally passes a WRDA
every two years to authorize new flood control and water management projects. However,
Congress has not updated the WRDA since 2000. In July of 2005, the House passed its version
of the WRDA update (H.R. 2864), well before Hurricane Katrina.50 The original House version
authorized $10-12 billion in spending on new water projects, including environmental restoration
on the Great Lakes, the Florida Everglades, and coastal wetlands in Louisiana.
The Senate took up the House bill on July 19, 2006, preserving most of the projects but
emphasizing some critical reforms. Senators McCain and Feingold successfully added an
amendment (S.Amdt. 4681) to create an independent peer review office for evaluating flood
control, navigation and environmental restoration projects costing $40 million or more.51 The
peer-review panels would address all elements of the projects, including cost, engineering and
design requirements, and environmental impact. They would include scientific and economic
experts responsible for making formal recommendations to the Corps. A decision to ignore the
panel's advice could be used against the Corps in legal proceedings.
47
Nat’l. Research Council, Review Procedures for Water Resources Project Planning (2002).
48
See also David Hosansky, Reforming the Corps, 13 CQ RESEARCHER (May 30, 2003).
49
Available at http://thomas.loc.gov/cgi-bin/bdquery/z?d109:S.728:.
50
Available at http://thomas.loc.gov/cgi-bin/bdquery/z?d109:H.R.2864:.
51
Available at http://thomas.loc.gov/cgi-bin/bdquery/z?d109:SP04681:. For discussion of the proposed legislation,
see Felicity Barringer, Senate Backs New Controls For Projects By Engineers, New York Times, July 20, 2006.
Brian Friel, Lawmakers Weighing Controversial Army Corps Reforms, National Journal, September 11, 2006
21
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Debate over the amendment explicitly centered on the Katrina disaster, the complicity of
the Corps, and the need for reform and independent oversight. Environmental groups, good
government organizations, and Katrina victims advocates all endorsed the changes. However,
several Senators in farm states with numerous Corps projects objected to the amendment,
arguing that it would lead to costly delays in approving projects. Led by Senators James Inhofe
(R-OK) and Kit Bond (R-MO), they offered their own amendment to water down the
independent review and preserve most of the House’s language on the issue (S.Amdt. 468252).
Ultimately, the Senate approved the McCain-Feingold amendment 54-46 and rejected the InhofeBond amendment 49-51
On September 19, the bill went before a joint House-Senate Conference Committee to
work out differences between the two chambers’ versions. The committee hopes to conclude
work by October 1. However, the Republican leadership named Senator Inhofe and
Representative Don Young (R-AK) as chairs of the Committee; both have been sharply critical
of efforts to increase oversight of the Corps. On the other hand, the White House supports the
Senate reforms. Therefore, chances are mixed whether the final bill will preserve the reforms.
Congress adjourned for the fall recess without reaching agreement on S.728/H.R. 2484, the
Water Resources Development Act that contained reforms of the Corps. Officials announced that
the conference committee would take up the bill again when Congress reconvened after the
midterms on November 13, but at this writing congressional inaction seems unlikely before the
end of the session. Given the change in control of Congress in the November 2006 elections, the
fate of the bill is even less predictable.
B. Rebuilding the Corps’ Technical/Engineering Capacity
The Corps’ engineering and R&D capabilities were degraded over the past twenty years
as a result of streamlining and budget cuts (downsizing and outsourcing). As a nation, we cannot
afford the loss of this expertise. Although outsourcing can be efficient in some instances, it
cannot be allowed to deplete the Corps’ own core expertise. The result of the outsourcing is to
separate the technical experts from the individuals making the ultimate decisions, creating a
gaping interstice where none needs to exist. As the National Research Council concluded,
“Shifting analytical tasks to the private sector, however, has its limits, as core, “in-house”
competence is necessary for the Corps to commission, manage, and comprehend the advice of
external experts.”53
The Army Corps of Engineers must be, first and foremost, the nation’s premiere expert in
flood control engineering. Through no fault of its own, the Corps has been stripped of much of
what it needs to perform this role. Congress must adopt a plan and allocate the necessary funds to
“put the ‘engineers’ back into the Corps of Engineers.” It must remake the Corps into the
organization that new, “wet behind the ears” civil engineers will want to join to sink their teeth
into their new profession. It must retain and perform sufficiently challenging engineering work to
encourage these engineers to develop their careers within the Corps. It must define and perform
sufficient R&D work to help support the activities of these engineers. And it must pay them
adequate salaries to be suitably competitive with private industry. An agency lacking in critical
technical expertise cannot be expected to engage in the kind of organizational learning so critical
to HROs.
52
Available at http://thomas.loc.gov/cgi-bin/bdquery/z?d109:SP04682:
53
NRC, supra note 46.
22
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
The Working Group for Post-Hurricane Planning for the Louisiana Coast has advanced
some complementary recommendations for Corps staffing in their report, A New Framework for
Planning the Future of Coastal Louisiana after the Hurricanes of 2005:
An essential element in enhancing the credibility and soundness of planning and
implementation is an agency's internal staff capabilities. The Corps of Engineers
is facing a significant loss of staff numbers and capability through retirement, just
at the time that the demands for its skills are increasing. Indeed, the integrated
planning process will demand a wider array of skills form the engineering,
hydrologic, geological, biological and social sciences than is currently available in
the agency or in federal or state agencies generally. Also, the effectiveness of the
long-term program requires the institutional memory that develops within a
permanent and professional staff.54
C. Restructuring the Federal/State Relationship in Flood Defense
The Corps’ relationship with local flood control entities in Louisiana is dysfunctional.
Some of the issues relate to the fragmentation of the local entities, which the state has begun to
address. However, a number of the issues are broader.
Often, water planning activities involve not only multiple federal agencies, but also state
and local governments. In the blunt words of one observer, “The first consequence is that flood
defense has no head . . . . Whatever the merits of this diffusion of authority, it does not produce
coherent flood control.”55 The result is an organizational structure that goes beyond having
“interstices” to having gaping chasms. Curing this problem is far from easy. One useful model
may be what has been called “modularity” -- a concept which involves provisional and
functional rearrangement of units in terms of alternative configurations of tools, structures and
relationships.56
The state of Louisiana has taken several useful steps to reduce administrative
fragmentation. On September 30, 2006, the voters overwhelmingly approved a measure to
consolidate the system of local levee boards. The amendment also requires levee boards to focus
on flood control rather than other distractions, and requires appointees to have relevant
expertise.57 This effort comes on top of Louisiana Governor Kathleen Blanco’s successful push
in November of 2005 to create a new interagency statewide flood control authority, the Coastal
Protection and Restoration Authority.58 The CPRA is charged with drawing up a master plan to
combine coastal restoration, hurricane protection, and flood control efforts at the state level, as
well as to provide oversight of local levee districts. Previously, statewide flood control programs
were scattered across different Cabinet positions and agencies. The CPRA also serves as the
single statewide “voice” of Louisiana when dealing with the Army Corps of Engineers and other
federal agencies.
54
WORKING GROUP FOR POST-HURRICANE PLANNING FOR THE LOUISIANA COAST, UNIV. OF MARYLAND
CTR. FOR ENVTL. SCI., A NEW FRAMEWORK FOR PLANNING THE FUTURE OF COASTAL LOUISIANA AFTER
THE HURRICANES OF 2005 (2006), ,http://www.umces.edu/la-restore.
55
Houck, supra note 14.
56
Jody Freeman & Daniel A Farber, Modular Environmental Regulation, 54 DUKE L.J. 795 (2005).
57
Frank Donze, Voters Merge Levee Boards, Times Picayune (Oct. 1, 2006).
58
Louisiana Coastal Protection, Conservation, Restoration, and Management Act, Act No. 8, 2005 La. Acts 1st
Extraordinary Sess. (codified at La. Rev. Stat. 49:213.1 et. seq.).
23
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Comprised of a 16-member panel that includes Cabinet secretaries, statewide officials,
and local levee officers, the CPRA will develop a comprehensive approach to protection that
includes central authority to plan, designed, build, inspect, and audit new flood control
measures.59 Governor Blanco has also directed the CPRA to design a formal state policy
regarding the Mississippi River Gulf Outlet, which overflowed during Katrina; create a levee
district inspection program and set minimum maintenance standards; and create a memorandum
of understanding that the state authority will sign with the Corps of Engineers that recognizes
that the CPRA will be the lead state entity to work with the Corps on hurricane protection and
coastal restoration.
Congress has directed the Army Corps’ New Orleans District to initiate a 24-month
endeavor, the Louisiana Coastal Protection and Restoration Project (LCPRA).60 The project will
identify, describe and propose a full range of flood control, coastal restoration, and hurricane
protection measures for South Louisiana. The LCPRA is currently in the midst of drafting its
first major reports. However, the LCPRA is advisory and shouldn’t be considered a new
“authority” in Louisiana. In the meantime, various federal authorities involved in the
reconstruction of New Orleans infrastructure (including the flood control system) lack any
organizational coherence.
D.. Developing a National Flood Defense Authority
A National Flood Defense Authority (NFDA) should be establish and charged with
oversight of the construction and maintenance of flood control systems. Following Louisana’s
lead, each state would have an equivalent organization that could foster cooperation and
development between and within the states. The Corps of Engineers, state flood control
authorities, and technical advisory boards would work with the NFDA to foster application of the
best available technology and help coordinate development and maintenance efforts and
planning. Federal and state governments would provide reliable and sustainable funding for the
life-cycle (design, construction, operation, maintenance) of specific flood defense systems. To
facilitate coherent funding, Congressional authorization and financing would be separated from
the traditional Water Resources Development Act process. Rather, the NFDA would be funded
through a separate appropriations process.
The Corps of Engineers, in cooperation with other qualified agencies and industrial
partners, would have the responsibility to design and construct, and if directed and authorized,
operate and maintain flood defense systems. The NFDA would be based on a continuous and
integrated process of flood risk assessment and management for specified flood defense systems,
with each of these systems being integrated with other allied flood defense systems. Flood risk
assessment and management processes would include proactive, reactive, and interactive
(adaptive) approaches based on the best available proven technology. Flood defense system
planning and development would engage public and industrial stakeholders and responsible
federal and state agencies in a cooperative and vigilant Technology Delivery System.
59
Governor Kathleen Blanco, Speech to the Coastal Protection and Restoration Authority, January 18, 2006,
available at http://www.gov.state.la.us/index.cfm?md=newsroom&tmp=detail&catID=4&articleID=1597.
60
Energy and Water Development Appropriations Act of 2006, Pub. L. No. 109-103, § 1003.
24
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Following the disastrous 1993 Midwest flooding, the federal Interagency Floodplain
Management Review Committee suggested similar reforms after evaluating the performance of
existing floodplain management programs. The Working Group for Post-Hurricane Planning for
the Louisiana Coast has advanced similar recommendations for organization and funding in their
report A New Framework for Planning the Future of Coastal Louisiana after the Hurricanes of
2005. This group observed:
Organizational and funding barriers that have inhibited the adoption of an
integrated planning and adaptive decision making process persist. Both new
organization and funding reforms are needed to support coastal planning and
project implementation by the Corps and the State.61
This group proposed a model that involves better federal intragovernmental coordination
with the new Louisiana Coastal Protection and Restoration Authority, and the
development of a Coastal Assessment Group and a Coastal Engineering and Science
Program. This model includes recommendations for programmatic authorization and
funding including formation of a new Louisiana Coastal Investment Corporation and
major revisions in the Water Resources Development Act appropriations process.
At the highest levels, we must also recognize that no one in the White House has the job
of disaster response. Yet federal disaster response requires action by many agencies – not just
FEMA but also the Defense Department, the Center for Disease Control, the Environmental
Protection Agency and others. White House coordination of these executive branch activities is
crucial. Just as the White House has a Council of Economic Advisors, it needs to have an official
or board charged with national disaster oversight – not necessarily day-to-day management
during disasters (which is FEMA’s responsibility), but rather a budgetary and policy supervision
role. This official would also be in charge of monitoring organizational problems in the agencies
charged with disaster prevention and response. Moreover, a natural part of the official’s portfolio
would be disaster prevention efforts, where the aim should be to avoid ever again being caught
unprepared for a “predictable surprise” like Katrina.
Such an integrated approach to catastrophic risk is lacking. One lesson from Katrina is
that disasters are not just engineering failures, they are social system failures and failures of
government. Societal infrastructures can collapse just as surely as physical ones can.
Consequently, disaster prevention cannot be considered in isolation from disaster response plans,
mechanisms for compensation and risk spreading, and reconstruction planning. All of these
issues are tightly coupled, yet the linkages receive little attention.
Under the Constitution, Congress bears the primary responsibility for developing national
policy and setting national priorities. Congress authorizes and controls FEMA, the Army Corps,
flood control projects, the flood insurance program, and other aspects of our nation’s response to
catastrophic risks. Yet Congress lacks the expertise needed to accomplish these tasks in a
systematic way. As a result, members of Congress operate as lone wolves, seeking to maximize
the benefits of individual projects for their districts, while Congress as a whole has no
mechanism for evaluating risk levels and project needs.
61
WORKING GROUP FOR POST-HURRICANE PLANNING FOR THE LOUISIANA COAST, supra note 30.
25
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
In the Post-Katrina Emergency Management Reform Act of 2006,62 Congress took a step
toward ensuring that it would be provided with better risk information. Section 649 calls upon
the Administrator of FEMA to assess on an on-going basis the country’s prevention capabilities.
But this mandate has two faults. First, the primary emphasis of sections 649-652 is on response
readiness, not on prevention. And of course, FEMA’s primary mission is to respond to
emergencies, not to assess the strength of levees or other preventative measures. Second, since
FEMA is an executive agency, it will necessarily have lower credibility within Congress than an
agency controlled by Congress itself. Thus, a Congressional Risk Office would be a significant
improvement.
V. Recommendations – Organizing for Success
Mobilizing the political will to build an adequate New Orleans Flood Defense System
(NOFDS) is the primary challenge for policy-makers. If the United States decides that Katrina’s
catastrophic effects must not be repeated, the necessary leadership, organization, management,
resources, and public support must be marshaled. One of the primary challenges is time; the
clock is ticking for this area of the United States to again confront a severe hurricane.
Our overall vision flood control is diagramed in the Appendix. We do not advocate a
completely centralized bureaucracy to manage flood control. Flood control must involve the
political input of both state and federal actors, because the ultimate decisions about tradeoffs
must be made in a politically accountable process. Moreover, flood control issues may require
multiple sources of expertise, not just the informed judgments of a single bureaucratic hierarchy.
What we have in mind, then, has some resemblance to what has been called “modular
regulation.”63
In its ideal, modularity supposes that both the tools and governance structures for
environmental regulation and resource management can be built, un-built and rebuilt—an
optimistic but, we hope, compelling normative view. Thinking in this way has both a theoretical
and practical payoff. From a theoretical perspective, modularity captures a moment of maturation
in both administrative law and environmental law, which has yet to be named and fully
described. In this moment, traditional forms of action and institutional structures give way to a
“problem-focus” that calls for new arrangements. We are trying to articulate what that moment
looks like, even as it is happening. From a practical perspective, we think modularity offers the
potential for real, measurable improvements in environmental policy-making. We believe it can
generate better-informed, more adaptive, and sufficiently accountable decisions that wind up
satisfying more stakeholders more of the time.
In essence, modularity is a recognition of organizational interdependencies and of the
need to insure effective information flow and learning across organizational interstices. Because
different problems require different alignments of expertise and involve different stakeholders,
modular structures are flexible and capable of being reassembled in new configurations.
To summarize, our specific recommendations are as follows:
Recommendation 1: Seriously consider defining risk within the framework of federal,
state, and local government responsibilities to protect citizens. For a technological delivery
system to enjoy legitimacy in a modern democracy, the risk levels that the system tolerates (and
62
Title VI of P.L. 109-295.
63
Jody Freeman & Daniel A. Farber, Modular Environmental Regulation, 54 Duke L.J. 795 (2005).
26
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
the corresponding margins of safety) must be set transparently. Environmental law provides a
number of models for making government decisions about acceptable levels of risk. One
possibility would be cost-benefit analysis; another would be to set (at least as an aspirational
goal) a fixed level of natural disaster risk considered acceptable.
Recommendation 2: Exploit the major and unprecedented role that exists for citizens to
shape this discourse. Citizen participation is a necessary part of governance because those who
govern do so at the informed consent of the governed. This includes the population exposed to
catastrophic risks, and the people that will be protected by the NOFDS. Authorities for
catastrophic risk management should ensure that those vulnerable have sufficient and timely
information regarding their condition and a reciprocal ability to respond to requests for their
informed consent especially regarding tradeoffs of safety for cost. The public protected by the
NOFDS need to be encouraged to actively and intelligently interact with its development.
Recommendation 3: Upgrade the Corps’ technical capacity. Congress must intensify,
focus, and fund Corps of Engineers modernization efforts; increase in-house engineering
capabilities and project performance; increase in-house research and development capabilities;
increase in-house engineering performance of technically challenging projects; develop an
organizational culture of high reliability founded on existing cultural values of duty, honor, and
country; develop a leadership role and responsibility for technical and management oversight of
all phases of development of a NOFDS. The Corps must re-establish technical superiority.
Outsourcing must be balanced with in-sourcing to encourage development and maintenance of
superior technical leadership and capabilities. This will require close and continuous
collaboration of federal legislative, executive, and judicial agencies. This will require that the
Corps of Engineers re-conceptualize itself as a pivotal part of a modular organization developing
partnerships with other federal agencies, state and local governments, enterprise interests, and
private stakeholders.
Recommendation 4: Restructure federal/state relationships in flood control. Enhance
cooperation and collaboration, reducing confusion as to overlapping areas of operation and
responsibility. Advance mutually supportive cross-checks and communication. In short,
recognize interdependency and work to ensure information flow across interstices.
Recommendation 5: Develop a National Flood Defense Authority (NFDA) charged
with oversight of the design, construction, operation and maintenance of flood control systems.
Each state would have an equivalent organization that could foster cooperation and
developments between and within the states. The Corps of Engineers, state flood control
authorities, and technical advisory boards would work with the NFDA to foster application of the
best available technology and help coordinate development and maintenance efforts and
planning. In cooperative developments, federal and state governments would provide reliable
and sustainable funding for the life-cycle of specific flood defense systems. This development
should be accompanied by development of an integrated and coherent Louisiana Flood Defense
Authority representing state, regional, local, city, and public stakeholders that can focus and
prioritize stakeholder interests and requirements and collaborate with the Corps of Engineers in
development of a NOFDS.
Recommendation 6: A new Council for Catastrophic Risk Management should be
appointed within the White House and given policy and budgetary oversight of disaster
preparation and response efforts. A similar body should be appointed within Congress,
analogous to the Congressional Budget Office. Incentives must be created to encourage all levels
27
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
of government to deal proactively and effectively with potential national, regional, and local
catastrophe.
Beyond these specifics, there is a larger lesson. In retrospect, while the New Orleans
disaster was triggered by a natural event, but that natural event caused unprecedented catastrophe
only because the institutional structures for designing and managing the flood control system
were grossly inadequate. Without these institutional failures, Hurricane Katrina would have
caused serious damage on the Gulf Coast and in New Orleans, but it would not have virtually
destroyed a major American city. The failures were found not only within the Army Corps of
Engineers, but in its relationships with outside institutions such as the White House and
Congress. Moving around boxes on an organizational chart has limited utility unless we fullya
assimilate the import of the historical record: Do not be mislead by years of quiet, “normal”
behavior. Disaster is always lurking, just around the corner. Natural disaster cannot be
prevented, but we can take appropriate precautions if our leaders and our organizations keep
their wits about them. Sloppy, stagnant organizations are merely wasteful in the short run; in the
long run, they can be catastrophic.
28
Farber et al., Reinventing Flood Control
Tulane Law Review
12/10/2006
Appendix: Proposed Flood Control Organization
Congressional
Catastrophic
Risk Office
Congress,
White House,
National
White House
Disaster Preparation
and Response Office
Outside review for
the Corps projects
National
Flood Defense
Authority
Army Corps
of Engineers;
Levees
and Wetlands
Local
agencies:
flood control,
environment,
and land use
Renewed
engineering
capacity
State flood
control and
environmental
agencies
Environmental
groups, local
communities,
Businesses,
Media
29